Background & myths · Knowledge
Does nofollow save crawl budget?
The idea sounds economical: nofollow the links into unimportant corners, and the crawler spends its visits on the pages that matter. Reality disagrees on both ends – and most sites never had a crawl-budget problem to begin with.
Since the 2020 hint era a nofollow link may be crawled anyway; and URLs are discovered through many paths – sitemaps, other links, history. A hint on one path steers little.
Who actually has the problem – and the tools that actually solve it
Calibration first: crawl budget – the number of URLs a search engine fetches from your site per period – becomes a practical concern for very large sites (think hundreds of thousands of URLs) and for shops whose filter combinatorics generate near-infinite parameter spaces. A typical company site, blog or plugin site is crawled comfortably in full; “optimizing” its budget optimizes a non-shortage. Why nofollow steers so poorly even where the concern is real: the attribute is a hint, one discovery path among many, and it does nothing for URLs the crawler already knows – a parameter jungle stays known forever. The tools that actually manage crawling, in working order: robots.txt disallow rules stop fetching of whole path and parameter patterns outright (the one true “do not crawl” instrument – noting that a disallowed URL can still appear in the index as a bare reference); clean information architecture avoids generating the jungle in the first place – curated filter pages instead of every combination, canonicals collapsing variants; noindex removes pages from the index where that is the actual goal (crawled, then dropped – the neighbouring meta-robots question sorts the pairs); and a current sitemap plus fast responses tell the crawler where the valuable pages are and let it fetch more of them per visit. In that toolbox nofollow simply has no crawl job – its job is the vouching question, and it does that one well. Diagnosis before therapy, as always: the search console's crawl statistics show what is actually fetched; “too much crawling of nonsense URLs” is a robots-and-architecture finding, not an attribute finding.
Key facts
- Scope check: crawl budget concerns very large sites and parameter-heavy shops – not typical sites.
- nofollow steers poorly: hint status, many discovery paths, known URLs stay known.
- True crawl control: robots.txt disallow for path/parameter patterns.
- Prevention beats steering: architecture, curated filter pages, canonicals.
- Diagnosis first: the search console's crawl stats show what is actually fetched.